concerning - dm5migu4zj3pb.cloudfront.netdm5migu4zj3pb.cloudfront.net/manuscripts/100000/... ·...

CONCERNINGTHE SPECIFIC RESPONSEOF GUINEA PIG'SRETICULOCYTESTO SUBSTANCESEFFECTIVE

IN PERNICIOUS ANEMIA'

By LOUIS S. GOODMAN,ARTHURJ. GEIGERAND THEODOREG. KLUMPP

(From the Departments of Pharmacology and Toxicology, and Internal Medicine,Yale University School of Medicine, New Haven)

(Received for publication February 24, 1936)

Since the value of liver and stomach prepara-tions in the treatment of pernicious anemia wasfirst recognized, attempts have been made toevaluate the potency of the effective materials.Clinical trial, which is the only generally acceptedmethod for testing these products (1, 18), is seri-ously handicapped by a growing scarcity of pa-tients suitable for assay purposes. Bioassaymethods employing laboratory animals have notproven successful. Forty-one such endeavorswere listed in a recent review (9). Much in-terest, therefore, centers about the reports ofJacobson and coworkers (7, 8, 9, 10) that thenormal guinea pig may serve as a valid indicatorof the therapeutic potency of materials effectivein pernicious anemia. Although the use of anormal animal appeared to violate a fundamentalphysiological principle of replacement therapy,it was deemed worth while to repeat Jacobson'sexperiments. Having an accumulation of experi-mental liver extracts awaiting potency determina-tions, we were eager to confirm the method andapply it to our needs. On the other hand, shouldthe method prove unsatisfactory, an early recog-nition of this would be desirable. Reports on thepotency of various chemically isolated fractionsderived from liver and assayed by Jacobson'sguinea pig method have already appeared in theliterature (3, 20). From experiments based onthis method, highly theoretical conclusions are alsobeing drawn concerning the relation of congo redto the pathological physiology of pernicious ane-mia (13), and the reaction of normoblastic bonemarrow to liver extract (11). Finally, thedanger inherent in the use of an unreliable bio-assay method by pharmaceutical firms is obvious.

' Aided in part by a grant from the Committee on Sci-entific Research of the American Medical Association.

METHODSANDPROCEDURE

The method of Jacobson, as outlined in hislatest publication (9), was followed precisely inthe great majority of our experiments; all devia-tions will be specifically noted. In order that thefollowing experiments may be clearly understoodsufficient details of Jacobson's technique will begiven.

Selection and care of animals. One hundred and eight-een normal adult guinea pigs weighing from 300 to 800grams were used. About one-third of the animals wereobtained from several local sources, and the remainderwere received from reputable breeders in other cities.For convenience, we chose to work chiefly with maleguinea pigs, as did Jacobson; only 20 were females.Wire cages with large-mesh screen bottoms housed from4 to 8 tenants each. The diet consisted of oats, carrotsand lettuce, as Jacobson recommended. In earlier experi-ments on 44 guinea pigs, cabbage was used in place oflettuce. A small number of animals, in an experiment tobe described later, was placed 6n a diet producing blacktongue in dogs.

Reticulocyte countintg. In some preliminary experi-ments the "permanent preparation" method of Osgoodand Wilhelm (17) was employed. This required about0.05 cc. of blood obtained from a minute ear vessel; thinand well-stained preparations were easily made. In thegreat majority of experiments, however, the simpler"wet preparation " technique employed by Jacobson wasfollowed.

Since we regarded a count of only 500 red blood cells,as employed by Jacobson, to be uncertain in accuracy forlow reticulocyte ranges, our reticulocyte estimations werebased on no less than 1000, and frequently 5000 to 7000cells. The counts were made by individuals experiencedin the technique. Each preparation was treated as an"unknown," and it was a frequent practice for the ob-servers to exchange slides for mutual checking. The re-liability and uniformity of the results were thus insured.

Method of administering liver extracts. All solutions,with either water or normal saline as the diluent, weremade up to a volume of 5 cc. and given intraperitoneallyin a single dose. Lederle's Solution Liver Extract Paren-teral, N.N.R., was used. Jacobson's basic dose wasadopted, namely, the extract derived from 4.3 mgm. offresh liver per kgm. of guinea pig. In most experiments,

435

LOUIS S. GOODMAN,ARTHURJ. GEIGER AND THEODOREG. KLUMPP

however, 10, 100 and 240 times this amount was injected.Control injections. In addition to potent liver extracts,

two control materials were employed: sterile physiologicalsaline solution and inactivated liver extract. Inactivationwas accomplished by heating the liver solution to above2000 C. for 30 minutes, and boiling the resulting charredmaterial for 4 hours. A third control procedure con-sisted in daily reticulocyte counts on uninjected animalsfor extended periods.

Selection of reactive guinea pigs. The assay methodof Jacobson is based on the use of so-called "reactive"guinea pigs; these animals are said to exhibit a specificreticulocyte response to liver extract. The criteria fortheir selection, and the fundamental characteristics oftheir reticulocyte behavior, as described by Jacobson, maybe briefly summarized as follows:

1. If the reticulocyte values of an unselected group ofnewly acquired adult normal guinea pigs are followed foran initial test period of 8 days, it is stated that the largemajority will consistently show levels below 1.2 per cent.Animals with reticulocyte percentages higher than 1.2, orthose which do not fall to this level or lower and remainthere for the last 4 days of this preliminary test period,are discarded as " unstable."

2. From 30 to 70 per cent of these stable animals aresaid to exhibit a reticulocyte response to potent liver ex-tract. These guinea pigs comprise the group known as"reactive." This means that at some time within the 6days following the injection, the reticulocytes should riseto 2.0 per cent or over for 2 consecutive days. Only ani-mals that are reactive in this manner to the initial testdose of liver are retained for further use as assay sub-jects.

3. Guinea pigs once reactive are said to be permanentlyreactive to the injection of potent liver extracts, and byvirtue of this remain suitable indefinitely for assay pur-poses.

Henwatopoietic response to liver extracts. Inasmuch asstable and reactive guinea pigs, as defined above, are saidto remain stable and to maintain indefinitely their specificreticulocytic responsiveness to liver with an approxi-mately constant degree of sensitivity, Jacobson uses thereactive animals in the assay of unknown liver prepara-tions. Furthermore, quantitative determination of po-tency is believed possible, since it appears that there is athreshold value for the response; and Jacobson placesthis minimally effective dose of potent liver at the amountof extract derived from 0.6 mgm. of fresh porcine liverper kgm. of guinea pig. This amount is termed a guineapig unit (G.P.U.). When a reactive guinea pig is givenone G.P.U. or any multiple of this, the reticulocytes willrise within 6 days to 2.0 per cent or over for 2 consecutivedays. It is stated that this response is all or none, andbears no relation to the amount of liver injected abovethe minimally effective dose; nor is there any other recog-nizable hematopoietic change of significance. In the as-say of an unknown product, Jacobson accepts a positiveresponse in 2 out of 3 injected test animals as evidenceof the presence of the anti-pernicious anemia substance.By testing various known dilutions, the potency of the

material can be quantitated. The reactive animals, aftera suitable rest period, are used over and over again; oneanimal was used 18 times in 21 months by the author ofthe method.

Fundamental postulates of the method. It isobvious that in order for the guinea pig assaymethod to be valid, the two following fundamentalcriteria must be fulfilled unequivocally:

1. A "stable " animal must remain stable.2. A "reactive " animal must show a reticulo-

cyte response only to potent anti-anemia prepara-tions, and not to indifferent substances; and itmust remain reactive.

If either or both of these postulates cannot besubstantiated, the method in our opinion is un-satisfactory.

RESULTS

To test the two basic tenets set forth above,1977 reticulocyte counts were performed in 260experiments on 118 guinea pigs.

Selection of " stable " guinea pigs. Our resultsshow an extreme variability in reticulocyte levelsof normal adult guinea pigs when observed for aneight-day period. This variability was observednot only in individual animals, but also in groupsof guinea pigs. For example, in one group of 12guinea pigs, 10 were stable on preliminary testing,but in another batch of 24 animals, only 1 wasfound stable.

When all our results on preliminary stabilitytests were analyzed nearly 48 per cent of all guineapigs had reticulocyte levels which rendered themunsuitable in the sense described above. This fig-ure is greatly in excess of that reported by Jacob-son (9), who claimed that the large majority ofall animals were stable on first trial. Regardlessof this discrepancy, the question immediately arosewhether the stable group, representing 52 per centof our guinea pigs, would remain stable. Wetried to answer this question in two ways. First,liver injection was withheld from a group of stableanimals, and their reticulocytes were followeddaily for from 8 to 20 days. Secondly, instead ofdiscarding those found unstable on first trial, alarge group was followed through second andthird observation periods. In this manner it hasbeen possible to learn much concerning the reticu-locyte fluctuations occurring spontaneously in nor-mal guinea pigs.

436

RETICULOCYTOSIS IN GUINEA PIGS

In regard to the so-called stable group, 25 ani-mals which met the requirements on first trialwere allowed to rest for one or more weeks, andthey were then re-examined for the stability oftheir reticulocyte levels. Thirteen were found tohave more than 2.0 per cent reticulocytes for 2consecutive days. This is at variance with theresults of Jacobson (9), who reported that hisstable group did not exceed 1.6 per cent reticulo-cytes when retested. If these animals had beeninjected with liver extract, these spontaneous re-

ticulocyte rises might have led us to believe that

I st Stability Tesf

c

O4-

a

1t-

0

0

over 50 per cent of the guinea pigs had respondedpositively. Typical charts of these experimentsare shown in Figures 1 and 2, which illustrate a

mild and an extreme instance, respectively, ofspontaneous fluctuations of reticulocytes in pre-

viously stable animals.It is obvious from these experiments that guinea

pigs show spontaneous reticulocyte fluctuationsof such a character that they cannot be dependedupon to remain stable simply because they hadhad low reticulocyte counts during one trial pe-

riod. Indeed, our results indicate that slightly

2nd Stability Test

0 2 4 6 8 10 0 4 8 12DAYS

FIG. 1. GUINEA PIG NUMBER17, d, 580 GRAMs. MILDSPONTANEOUSRETICULOCYTE FLUCTUATION IN A PREVIOUSLYSTABLE ANIMAL.

C-w

oi

4w

vL

._

c.

DAYS

FIG. 2. GUINEA PIG NUMBER25, d, 520 GRAMS. EXTREMESPONTANEOUSRETICULOCYTEFLUCTUATION IN A PREVIOUSLY STABLE ANIMAL.

437


1sf Stability Test 2nd Stability Test

3 5DAYS

3rd Stability Test

FIG. 3. GUINEA PIG NUMBER50, d, 740 GRAMS. SPONTANEOUSRETICULOCYTEFLUCTUATIONAFTER Two PREVIOUS STABLE PERIODS.

over 50 per cent of all guinea pigs stable duringone observation period will be found unstable ifre-examined at a later time. Furthermore, ifone now takes those animals remaining stablethrough a second trial period and tests them a

third time, again an appreciable number will befound unstable. For example, in one group of11 guinea pigs stable during two previous obser-vation periods, 5 were no longer stable when sub-jected to a third trial. In Figure 3, this occur-

rence is illustrated, and the experiment depictedis typical of 4 others.

Repeated Gontinuous

Ist

c

a-

>)

C)0

75

From these experiments it is clear that in theguinea pig reticulocyte values fluctuate spontane-

ously. The reticulocyte behavior during any

single period of observation, therefore, is not a

valid basis upon which to label an animal stable or

unstable. The particular result obtained dependsentirely upon what part of the reticulocyte curve

is being observed. If a stable guinea pig is sub-jected to repeated periods of observation, it willsooner or later be found unstable. This is wellseen in Figure 4, which illustrates continuous dailyreticulocyte counts taken on a stable animal over

; Stabiliiy Tests

2nd 3rd 4th

2 6 10 14 18 22 26 30

DAYSFIG. 4. GUINEA PIG NUMBER14, d, 540 GRAMS. SPONTANEOUS

RETICULOCYTE FLUCTUATIONS DURING PROLONGEDOBSERVATIONOF A

PREVIOUSLY STABLE ANIMAL.

438

-

c

4)0

0.-I

cc.


Ist Sfability Test 2nd Stability Test1 , , _

1 3 5 7 9 1 3 5 7 9DAYS

FIG 5. GUINEA PIG NUMBER98, d, 510 GRAMS. STABLE RETICULO-CYTE PERIOD IN A PREVIOUSLY UNSTABLEANIMAL.

a period of 20 days, and is typical of 5 similarexperiments.

Next, a study was made of the reticulocytebehavior of those animals which were unstable inthe first observation period. Instead of beingdiscarded, as Jacobson recommended, these ani-mals were allowed to rest and were then re-

examined. Of 29 such animals, 14 or slightlyless than 50 per cent were found to be stablewhen given a second trial. Of the remaining 15,which were again unstable during the second trialperiod, 10 were allowed to rest and then tested a

third time. Of these, 6 animals were now foundstable. In other words, only 4 animals of theoriginal group of 24 that were followed remainedunstable through three trials. It is probable thatif these 4 animals were subsequently retested

they too would ultimately have stable reticulocyteperiods. The importance of these experimentslies in the possibility that if by chance all initiallyunstable guinea pigs had been observed a fewweeks earlier or later, a large number would havebeen accepted as reliable test animals instead ofhaving been discarded. Figure 5 shows an ex-

periment on an unstable guinea pig, which, whentested three weeks later, was found stable; thisexperiment is typical of 13 similar ones. Fig-ure 6 illustrates that an animal, unstable duringtwo separate trial periods, was found stable onthe third testing; there were 5 experiments withsimilar results.

From the above evidence, it may now be saidthat not only is it impossible to rely on the reticu-locyte constancy of a so-called stable animal, but

4 Ist Stability rest

1 3 5 7 9

2 nd Stability Test

1 3 5DAYS

3rd Stability Test

1 3 I

FIG. 6. GUINEA PIG NUMBER105, d', 480 GRAMS. STABLE RETICULOCYTEPERIOD AFTER Two PRE-VIOUS UNSTABLEPERIODS.

439

c

0-

4)

4):1.

5-

`00

0.

16--- -- -Z. -- --- -

I m~~~~~~~~~~~~~V

2 0) ________4C 4)

co_ e s s 5 e 7 ffi C: 7

I

7 9


it is also impossible to discard a particular guineapig as unstable. The results, in our opinion, defi-nitely indicate that if one is guided by the criteriaof the method, all guinea pigs are unsuitable be-cause of their normal, spontaneous and unpre-

dictable reticulocyte fluctuations. It has not beenpossible, therefore, to substantiate the first postu-late of the method.

The specificity of the reticulocyte response toliver injections

I. Response to injection of liver. Despite themarked and unpredictable fluctuations of reticulo-cytes occurring in normal guinea pigs, we deemedit advisable to investigate the method further todiscover whether there might be some character-istic response to potent liver easily distinguishablefrom any spontaneous change in reticulocyte level.

After a short rest period, potent liver extractwas administered to 44 guinea pigs meeting thequalifications for stability. The experimentswere divided roughly into four groups given re-

spectively Jacobson's routine dose and also 10,100, and 240 times this dose. Positive responses

were obtained in 15 experiments, or 34 per cent.No difference was noted in regard to the amountof liver injected. The group receiving the small-est dose showed the same variability of reticulo-cyte behavior as the group receiving the largestdose. In Figure 7 is illustrated a positive andnegative experiment in the same animal which on

two occasions, separated by a suitable rest period,

Potent Liver

c

.-

a

_ A

co

1 3 5 7 9

received 240 times the routine dose of liver ex-

tract. A positive reticulocyte response appearedto follow the first injection; there was no ap-

parent effect from the second injection. A simi-lar result was obtained in 2 other experiments.

These experiments indicate that a guinea pigwhich has responded positively cannot be reliedupon thereafter always to respond positively toliver. Jacobson's contention, therefore, that " thereactive state is maintained indefinitely " was notconfirmed.

II. Response to injections of control sub-stances. Observations made on normal unin-jected guinea pigs constitute important controldata. Additional control experiments were per-

formed by injecting isotonic solutions of sodiumchloride and totally inactivated liver extract. Thelatter was given in doses corresponding to thoseused for the potent extracts. The method ofinactivation has been described.

Injections of salt solution were made in 27stable guinea pigs. Many of these animals hadundergone prior injections of liver or repeatedstability observations; the remainder received thesaline injection directly following the first stabil-ity test. Ten of these 27 experiments resultedpositively, and thus 37 per cent of the nonspecificsaline injections produced the type of response

reserved by the Jacobson method exclusively forpotent liver materials.

Injections of inactivated liver were made in 16stable guinea pigs, and 10 of these, or 63 per cent,

Potent Liver

1 3 5 7 9DAYS

FIG. 7. GUINEA PIG NUMBER18, d, 445 GRAMS. NEGATIVE RESPONSETOLivER INJECTION IN A PREVIOUSLY REACTIVE ANIMAL.

5

4- .2

C a.~~~~~~c 0. Q~~~~~~

0

#

2 7

~~~~~~~~~~.X

w

AA0


were positive. In other words, the majority ofthe animals reacted positively to impotent liver.

From these control results, it appears that hereagain one is observing only normal spontaneousfluctuations of reticulocytes, and that the injec-tions of liver extract or saline are irrelevant andmerely coincidental.

In Figure 8 is shown an example of a positiveresponse to the injection of saline solution.

Stability Test

1 3 5 7DAYS

Saline

1 3 5 7 9

FIG. 8. GUINEA PIG NUMBER54, c, 535 GRAMS. POSI-TIVE RESPONSETO SALINE INJECTION.

Figure 9 illustrates a positive response to theinjection of inactivated liver extract solution.

In Figure 10 is depicted a control experimentin a stable animal which had reacted positively tothe injection of liver. Then, to test the specificityof the liver response, saline and inactivated liverinjections were given. Positive reticulocyte re-sponses followed each of the control injections.This experiment is typical of 5 others.

Sfability Test Potent LiverS0 0

42)

.2~~~~~~~~~0

7 9DAYS

Stability Test Inactivated Liver

DAYS

FIG. 9. GUINEA PIG NUMBER48, e, 625 GRAMS. POSI-TIVE RESPONSETO INAcTIVATED LIVER INJECTION.

A comparison was made of the reticulocytecurves of all positive responses including those oc-curring spontaneously and after active liver, in-activated liver, and saline injections. The analy-sis included the character of the curves in regardto contour, the average height of reticulocytepeaks, the total reticulocyte response, the durationof the response, and the day on which the peakwas reached. No features were found whichwould distinguish the potent liver responses fromthe non-specific ones. The results are listed inTable I.

From the evidence presented in this section,therefore, it is clear that the second postulate con-cerning the specifiicity and permanency of thereticulocyte response to liver extract could not besubstantiated.

Inactivated Liver Saline

FIG. 10. GUINEA PIG NUMBER46, e, 580 GRAMS. POSITIVE RESPONSESTO INACTIVATED LIVER AND SALINE IN APREVIOUSLY STABLE AND REACTIVE ANIMAL.

30

-w2> 16

.2.2w

CZ

4.)

o Ul)L.. I

-

N __ <___V_<

e t e I

441

m4p4i1-4DCL.

40:1%4i

-C)4i

4pcr.

. . I I


TABLE I *

Analysis of positive reticulocyte responsesin stable guinea pigs

Reticu- Dura- Day of retitalTreatment locyte of re peak locyte

peak sponse responise responsefigure t

per cent days per centActive liver (15 ani-

mals) .............. 3.9 4 4th 2.9Saline (10 animals).... 3.5 4 4th to 5th 2.8Inactive liver (10 ani-

mals) ........ ..... 3.2 3 4th 2.6Uninjected (13 animals) 3.8 4 3rd to 4th 3.2

* All figures represent the average of the values for allexperiments in each group.

t This figure is obtained by averaging those reticulocytevalues over 1.8 per cent during the period of response.

The method, we believe, has established by def-inition an arbitrary division between stable andunstable guinea pigs, and positive and negativeresponses. It is our experience that this dividingline of 1.2 per cent reticulocytes, which differ-entiates between stable and unstable animals, isdrawn near the mean in the range of reticulocytefluctuations of normal guinea pigs. For example,if our interpretation is correct concerning thespontaneous character of these reticulocyte values,it would be of interest to know the range of dis-tribution of the reticulocyte counts obtained inall our experiments. This is shown in Figure 11,in which curve A represents the distribution of1977 reticulocyte counts done on 118 guinea pigsregardless of the various procedures followed inparticular animals. It is interesting that 1029counts, or 52 per cent of the total, are 1.2 percent or below. This correlates well with theprevious finding that approximately one-half ofour guinea pigs were stable.

In Figure 11, curve B represents the distribu-tion of the reticulocyte counts of all the potentliver experiments; curve C, similarly, representsthe distribution of the reticulocyte counts of allthe control experiments. There is no significantdifference in the character of these reticulocytedistributions.

It is merely chance that in any fairly largegroup of animals, regardless of various harmlessprocedures, a certain number will show reticulo-cyte values over 2 per cent for 2 consecutive days.The definition of a positive response is, in ouropinion, an arbitrary one.

25

20L

O 150

.00

I-.2

r-

-a

62

1o

5

A--AII Experiments

B- Liver Experiments .-.

C=Control Experiments ---

0O 0.4 Q8 2 L6 20Q3 0.7 1.1 1.5 l9 23

Reticulocyte Values

a4 28 322.7 3.1 365

FIG. 11. DISTRIBUTION OF 1977 RETICULOCYTECOUNTSREPRESENTINGALL EXPERIMENTS (CURVE A), AND COM-PARISON WITH 439 COUNTS IN LIVER EXPERIMENTS(CURVE B), AND 1538 COUNTSIN CONTROLEXPERIMENTS(CURVE C).

In Table II is presented a statistical summaryof all the experiments performed on stable guineapigs.

TABLE II

Statistical summary of results of experiments performedon stable guinea pigs

Positive NegativeNumber responses responses

Treatment experi-ments Num- Per Num- Per

ber cent ber cent

Active liver ........... 44 15 34 29 66Saline ................ 27 10 37 17 63Inactive liver .......... 16 10 63 6 37Uninjected ............ 25 13 52 12 48

Results in regard to sex, source and diet. Nodifferences were noted in groups of guinea pigsacquired from different sources. The data pre-sented by males and females were identical. The

442


results in the preliminary experiments with 44guinea pigs given cabbage as the green vegetablewere the same as the results with the 74 guineapigs fed lettuce; nor were there any changes whenthe cabbage-fed animals were later placed on let-tuce.

Hematological observations. In 23 animals themain features of the blood picture were studied.Red cell counts were done in duplicate, usingcertified blood pipettes and counting chambers.Hemoglobin readings were made on the Haden-Hauser hemoglobinometer calibrated by the oxy-gen capacity determination. Packed red cell vol-umes were measured with Van Allen hematocrits,employing 2 per cent potassium oxalate solution.The mean and extreme findings are presented inTable III.

TABLE III

Hematological data on 23 normal guinea pigs

| belcout Hemoglobin Hematocrit

millions grams per cent ofper cent packed ceUs

Average ............. 6.10 14.7 42.0Range. 7.04-4.73 16.0-13.0 47.0-37.0

These figures compare favorably with thevalues presented by Jacobson, and suggest thatour animals were not hematopoietically differentfrom his guinea pigs. We found no correlationbetween blood pictures, reticulocyte variations, andtreatment. The level of the circulating reticulo-cytes is usually regarded as a sensitive index ofbone marrow activity. The rise and fall of thesecells in our experiments were not correlated withany increased or decreased demand detectable inthe blood picture. This again suggests that thevariable reticulocyte values in these animals rep-

resent only the usual activity of the normal guineapig's marrow. It is not remarkable, then, thatin the absence of a specific hematopoietic de-ficiency, no characteristic response follows theinjection of liver extract.

Results in regard to reticulocyte countingmethod. When a comparison was made of thedata obtained by the " permanent preparation"method of Osgood and Wilhelm (17), which was

used in about 20 per cent of our experiments,and the " fresh preparation " method used in theremainder, no significant differences were noted.

The small daily loss of blood incidental to thereticulocyte estimations had no effect on the reticu-locyte values or blood count. This was deter-mined in 6 animals that were bled daily far inexcess of the amount of blood required for thereticulocyte counts.

In our opinion a reticulocyte percentage basedon a count of only 500 red blood cells is unreliablein these low reticulocyte ranges. Unusual carewas exercised to insure perfection in the prepara-tion of the blood films; nevertheless, the valuesobtained for duplicate reticulocyte counts onblocks of 500 cells each were often significantlydifferent. The necessity for counting at leasttwo blocks of not less than 500 cells each is seenin Figure 12 which illustrates a few of the morestriking instances in which, by counting only 500cells, one might conclude that the animal wasstable or non-reactive, and then by counting adifferent block of 500 cells the animal would bejudged unstable or reactive.

Wedo not feel that the presentation of thesedata labels our counting technique any the lessaccurate, but rather that it indicates a source ofdifficulty and possible error in a relatively simplehematological procedure even when done care-fully by experienced workers. It is our under-standing that the accepted standard for reticulo-cyte counting in clinical work is a minimum of1000 red blood cells for reticulocyte levels under5 per cent. Since Jacobson's method necessitatesthe differentiation between 1.5 and 2.0 per centof reticulocytes, permitting an error of less than3 reticulocytes per 500 red cells, it would seemeven more necessary to examine at least 1000 cells.

Reticulocyte behazi or of guinea pigs on theGoldberger diet producing black tongue in dogs.Inasmuch as we were unable to obtain truly stableguinea pigs, it was thought desirable to attemptto stabilize the circulating reticulocytes at a lowlevel and to render the bone marrow specificallysusceptible to liver extract. We sought to ac-complish this by feeding a deficient diet.. Theresults of some of these experiments (2) arereported here because they shed additional lighton the bone marrow activity of this species. Themany lines of evidence which suggested the useof the diet producing pellagra can be indicatedonly briefly here. In Goldberger's hands the diet

443

LOUIS S. GOODMAAN,ARTHURJ. GEIGEIR AND tHEODOREG. XLUMPP

FIG. 12. 19 REPRESENTATIVE INSTANCES OF SIGNIFICANT DIFFERENCES IN DUPLICATERETICULOCYTECOUNTS, OF 500 CELLS EACH, MADEON THE SAMEPREPARATION.

Note that in each instance the duplicate counts f all on opposite sides of the shaded" crucial zone " which defines reactivity and the upper limit of stability.

was productive of pellagra in humans (5).Rhoads and Miller (19) reported that this dietcaused anemia and black tongue in dogs, while inswine it produced a syndrome strikingly similarto human pernicious anemia (15). The diet hasbeen thought to be deficient in vitamin B2, andthe relationship between the anti-pernicious ane-mia principle and vitamin B2 has received con-siderable attention. Finally, Miller and Rhoadsfound that the anti-anemia substances, when fedorally, possessed a life-saving factor for guineapigs maintained on this diet (14).

Guinea pigs were fed the diet producing pel-lagra (4, Table 6) plus orange juice, but omittingsucrose, and their blood pictures were followed.In the majority of these animals the reticulocytesdecreased within two weeks to extremely lowlevels. In many instances no reticulocytes wereseen even on counting as many as 5000 to 10,000red blood cells, and counts of 0.02 per cent werenot rare. In 4 animals showing such extremelylow counts for from several days to a week, singleintraperitoneal injections of 0.25 cc. of Lederle'sconcentrated liver extract solution were given.The reticulocytes rose rapidly in 3 guinea pigs,more slowly in the fourth one, but in all instancesthere occurred what appeared to be a definite andmarked response with an increase of reticulocytesto from 2.0 to 2.5 per cent. As a control, there-

fore, another group of 4 animals on the same dietwas observed daily, and it was noted that thereoccurred spontaneous increases in reticulated cellsmuch greater than those observed after liver ex-tract. In 2 animals that lived on this diet fortwo months, several periods of alternating ex-tremely low and very high reticulocyte levels wereseen, and these phasic variations were relatedneither to any observable change in the well-being

3r

2

3 5

cI0

A I I a t- 1-

7 9 l1 13 15DAYS

17

FIG. 13. GUINEA PIG NUMBER110, d, 320 GRAMS.GOLDBERGERDIET. APPARENT POSITIVE RESPONSE TOLIVER INJECTION.

.44

c00

e-a)C36

0-0-.1%ti0

752>0

cr.


of the guinea pigs nor to the loss of weight whichoccurs on this regimen. Here again, then, onefinds marked fluctuations in animals which forshort periods of observation appear to have re-duced bone marrow activity and low and stablereticulocyte values induced by a deficient diet.Figure 13 illustrates an apparent effect of liverinjection given during a period of reduced bonemarrow activity, and Figure 14 presents a com-parable experiment in which liver injection waswithheld but the reticulocytes counted daily forover two months.

From these experiments it seems that this dietaltered bone marrow activity in such a mannerthat the swings in peripheral reticulocyte levelsseen in normal guinea pigs were markedly accen-tuated. The reticulocyte behavior on the deficientdiet differed not in kind but only in degree fromthat of normal animals. Thus, what appeared tobe a promising method for rendering the guineapig a suitable assay animal has failed in our hands.

It should be added that the daily oral admin-istration of large amounts of Lilly's Liver Ex-tract, N.N.R., to these guinea pigs had no char-acteristic effect on the reticulocytes. The responseobtained by Jacobson (9) from the oral admin-

8

7

6-

5

4

0.40

istration of anti-anemia substances to normalguinea pigs finds no support in these experiments.

DISCUSSION

The failure of our experiments to corroborateJacobson's assay method is clear from the resultspresented. This discrepancy cannot reasonablybe ascribed to an inherent hematological differ-ence between his animals and ours. We usednormal adult guinea pigs deliberately obtainedfrom several sources. The animals remainedhealthy, and the blood pictures in a representa-tive group were entirely normal throughout theexperiments.

Jacobson (9) suggested that the megaloblasticbone marrow which he observed in his reactiveanimals bore a close causal relation to their suit-ability for assay purposes, and stated that "thelarge number of megaloblasts in proportion tothe number of normoblasts, simulate the classicalpicture of the bone marrow findings in perniciousanemia." He found, however, the same type ofmarrow in the non-reactive guinea pigs, and fromthe remainder of the hematological data it isprobable that his finding is the normal conditionfor this species. A megaloblastic marrow oc-

FIG. 14. GUINEA PIG NUMBER114, el

30 40 50DAYS

535 GRAMS. GOLDBERGERDIET.RETICULOCYTELEVELS.

60 70

MARKEDSPONTANEOUSSWINGSIN

445


curring normally in a non-anemic guinea pig isobviously entirely different in significance fromthe abnormal megaloblastic hyperplasia seen inpernicious anemia patients in relapse. It may berecalled here that once previously the impressionthat a megaloblastic marrow in a normal animalrendered that species suitable for liver assay pur-poses resulted in much futile work on pigeons.2

It would appear at present that there is nophysiological basis for the development of a spe-cific reticulocytosis to liver in animals with normalblood, and in our experience the guinea pig is nodifferent in this respect from the other speciesof normal laboratory animals which have failedto yield a bioassay method for anti-perniciousanemia substances.

A search of the literature revealed only tworeports apparently confirming the usefulness ofthe normal guinea pig as a test animal. Thefirst study, by Landsberg and Thompson (12),was based on only 6 animals, and no controlswere mentioned. In 4 of their 6 animals thepre-injection reticulocyte levels would render themunsuitable for the method here under considera-tion. The report of Miller and Rhoads (16) de-serves special mention because these workers im-plied confirmation of Jacobson's results and statedthat " increases in the number of circulating re-ticulocytes which are similar in degree, duration,and time of occurrence to those obtained by Jacob-son may be induced by potent, anti-anemic sub-stances which are orally as well as parenterallyadministered." A careful analysis of their data,however, reveals that their results differ fromJacobson's in important respects. In the firstplace, the height of reticulocytosis obtained byMiller and Rhoads was definitely greater thanJacobson's. The latter's positive responses wereunder 3.4 per cent in 90 per cent of 193 experi-ments, whereas, in Tables II and III presented byMiller and Rhoads, at least 5 responses out of the11 recorded were over 3.4 per cent, and in one

2 It is of interest to note that very recently Jones (11)investigated the bone marrow of the guinea pig, pigeon,and rat. He found that the red cells of these marrows" belong to the definitive or normoblastic series, and theyare identical with those of the marrow of the normalhuman adult, which does not possess megaloblasts."There was " -no morphologic evidence for a megalo-blastic bone marrow in the guinea pig," and no similarityto the megaloblastic marrow of pernicious anemia.

experiment all 3 animals reached 5 per cent orover. Since the method deals essentially withlow reticulocyte values, such a difference in oneof its fundamental phenomena appears signif-icant. More important, however, is the fact thatMiller and Rhoads' reticulocyte responses did notoccur until the seventh to thirteenth day after theinitiation of treatment, while Jacobson's criterionof a positive response demands that it occurwithin six days. The importance of this differ-ence is well illustrated in Jacobson's Table VII(9) which shows that animal 256 did not respondwithin six days to a presumably impotent materialgiven orally. This animal was then immediatelygiven a second mixture, and this time a reticulo-cyte response did occur within the required periodof six days. On the basis of these time relation-ships, this so-called positive response, coming asit did on the twelfth day after the original im-potent material, was regarded by Jacobson asevidence of potency of the second and impotencyof the first preparation. Miller and Rhoads, onthe other hand, might readily have regarded thefirst preparation as potent, because they acceptedsuch late reticulocyte responses as positive. In-deed, inasmuch as all the positive responses ob-tained by Miller and Rhoads occurred after thefirst week, these same positive responses would,if judged by Jacobson's criteria, be called negativeresponses.

These discrepancies in the degree and time ofappearance of the reticulocytosis cannot be as-cribed to differences in dosage. Jacobson (9,Table V) noted no influence on the magnitude ofthe reticulocyte rise after the injection of theminimal effective dose as compared with 6000times this quantity. Furthermore, he noted atendency for the responses to occur earlier, ifanything, within the six-day observation periodwhen larger amounts of liver were given. Ourown interpretation of these data is that hereagain one is encountering nothing more than theusual non-specific and spontaneous reticulocytefluctuations of normal guinea pigS.3

3 In a recent personal communication C. P. Rhoadshas written, "Further experimentation has advancedincontrovertible evidence that guinea pig reticulocytesmay increase to a level of over 2 per cent from a varietyof causes unassociated with the administration of sub-stances effective in the treatment of pernicious anemia.Experiments are under way in an attempt to elucidate

446


We have carefully observed the four precau-tions listed by Jacobson in his discussion (9) inregard to the diet and environment of the pigs, thehigh degree of accuracy in reticulocyte counting,the selection only of animals stable during a pre-liminary control period for liver injection, andthe use of liver extract of a high degree of thera-peutic potency. Despite meeting these and theother criteria of the method, on the basis of ourexperience we are forced to conclude that thenormal guinea pig exhibits, in respect to his re-ticulocytes, such a capricious behavior that theuse of this species for bioassay of liver extractis at present unsuitable and hazardous.

The prediction is ventured that a satisfactorylaboratory assay animal will not be obtained untilit is possible to produce experimentally a clinicalpicture closely simulating the pernicious anemiaof man. Whether the anemia of swine producedby Miller and Rhoads (15), or the anemia asso-ciated with liver cirrhosis obtained in rats byHiggins and Stasney (6), will provide the long-sought laboratory test animal remains a questionfor further experimentation.

CONCLUSION

The normal adult guinea pig shows considerableand unpredictable spontaneous fluctuations in re-ticulocyte levels. These variations are of such anature as to render this normal animal unsuitablefor assaying the potency of materials effective inpernicious anemia.

The authors are grateful to Mrs. Helen Good-man for her invaluable technical assistance.

BIBLIOGRAPHY1. Dameshek, W., and Castle, W. B., Assay of com-

mercial extracts of liver for parenteral use inpernicious anemia; method of successive reticulo-cyte responses in the same patient. J. A. M. A.,1934, 103, 802.

2. Doeller, C., Goodman, L. S., and Geiger, A. J., (Ex-periments to be published).

3. Fiske, C. H., Subbarow, Y., and Jacobson, B. M., Themultiple nature of the pernicious anemia principlein liver. J. Clin. Invest. (Proc.), 1935, 14, 709.

4. Goldberger, J., Wheeler, G. A., Lillie, R. D., andRogers, L. M., A further study of butter, fresh

the cause of spontaneous variations in numbers of reticu-locytes; such spontaneous variations demand that thegreatest care be used in employing the guinea pig as atest for anti-anemic substances."

beef, and yeast as pellagra preventives, with con-sideration of the relation of factor P-P of pellagra(and black tongue of dogs) to vitamin B. U. S.Pub. Health Rep., 1926, 41, 297.

5. Goldberger, J., and Wheeler, G. A., The experimentalproduction of pellagra in human subjects by meansof diet. Nat. Inst. Health Bull., 1920, No. 120, 7.

6. Higgins, G. M., and Stasney, J., Macrocytic anemiain experimental cirrhosis. Proc. Staff Meet., MayoClinic, 1935, 10, 429.

7. Jacobson, B. M., On a pernicious anemia-like state inthe guinea pig. J. Clin. Invest. (Proc.), 1934, 13,714.

8. Jacobson, B. M., The response of the normal guineapig to the administration of liver extracts. Sci-ence, 1934, 80, 211.

9. Jacobson, B. M., The response of the guinea pig'sreticulocytes to substances effective in perniciousanemia; a biologic assay of the therapeutic po-tency of liver extracts. J. Clin. Invest., 1935, 14,665.

10. Jacobson, B. M., The assay on guinea pigs of thehematopoietic activity of human livers, normal andpernicious anemia. J. Clin. Invest., 1935, 14, 679.

11. Jones, 0. P., The reaction of normoblastic bone mar-row to liver extract. J. Lab. and Clin. Med.,1936, 21, 335.

12. Landsberg, J. W., and Thompson, M. R., The guineapig as a hematopoietic test animal. J. Am. Pharm.A., 1934, 23, 964.

13. Mermod, C., and Dock, W., A colloidal dye effectivein treating pernicious anemia and evoking reticulo-cytosis in guinea pigs. Science, 1935, 82, 155.

14. Miller, D. K., and Rhoads, C. P., Absence of dietaryanti-anemia substance in the diet causative of ca-nine black tongue. Proc. Soc. Exper. Biol. andMed., 1934, 32, 419.

15. Miller, D. K., and Rhoads, C. P., The experimentalproduction of loss of hematopoietic elements of thegastric secretion and of the liver in swine withachlorhydria and anemia. J. Clin. Invest., 1935,14, 153.

16. Miller, D. K., and Rhoads, C. P., The reticulocyteresponse in guinea pigs following the oral admin-istration of certain anti-anemic substances. NewEngland J. Med., 1935, 213, 99.

17. Osgood, E. E., and Wilhelm, M. M., Reticulocytes.J. Lab. and Clin. Med., 1934, 19, 1129.

18. Pharmacopoeia of the United States of America, 11thDecennial Revision, Mack Printing Co., Easton,Pa., 1935, p. 484.

19. Rhoads, C. P., and Miller, D. K., The production indogs of chronic black tongue with anemia. J.Exper. Med., 1933, 58, 585.

20. Subbarow, Y., Jacobson, B. M., and Fiske, C. H., Theseparation of the substances in liver which arereticulocytogenic in the guinea pig and which aretherapeutically effective in experimental canineblack tongue. New England J. Med., 1935, 212,663.

447

concerning - dm5migu4zj3pb.cloudfront.netdm5migu4zj3pb.cloudfront.net/manuscripts/100000/... ·...

Documents